Functional Connectivity’s Degenerate View of Brain Computation

• DOI: 10.1371/journal.pcbi.1005031
• Publication Date: 2016-10-13T18:08:31Z
• Authors: Guillaume Marrelec; Arnaud Messé; Alain Giron; David Rudrauf
• Keywords: Representation

Brain computation relies on effective interactions between ensembles of neurons. In neuroimaging, measures functional connectivity (FC) aim at statistically quantifying such interactions, often to study normal or pathological cognition. Their capacity reflect a meaningful variety patterns as expected from neural in relation cognitive processes remains debated. The relative weights time-varying local neurophysiological dynamics versus static structural (SC) the generation FC measured unsettled. Empirical evidence features mixed results: little significant variability and correlation with functions, within participants. We used unified approach combining multivariate analysis, bootstrap computational modeling characterize potential SC both qualitatively quantitatively. data simulations generative models different dynamical behaviors demonstrated, largely irrespective metrics, that linear subspace dimension one two could explain much across FC. On contrary, BOLD time-courses not be reduced small subspace. appeared strongly partly governed by Gaussian process. main differences simulated empirical related limitations DWI-based estimation (and itself then estimated FC). Above beyond limited range signal itself, may offer degenerate representation brain access underlying complexity. They feature an invariant common core, reflecting channel network conditioned SC, limited, though perhaps residual variability.